US2024138793A1PendingUtilityA1

Estimating pharmacokinetics of a contrast medium through a cardiovascular system analytical model and/or neural network algorithm

Assignee: SIEMENS HEALTHCARE GMBHPriority: Oct 31, 2022Filed: Oct 30, 2023Published: May 2, 2024

Est. expiryOct 31, 2042(~16.2 yrs left)· nominal 20-yr term from priority

Inventors:Pooyan Sahbaee Bagherzadeh Ralf Gutjahr Alexander Muehlberg Marcel Reimann Ehsan Samei

G06F 2119/14G06F 2113/08A61B 6/54A61B 6/52A61B 6/504A61B 6/035G06F 30/27G06F 30/28G16H 30/00G16H 50/70G16H 50/50A61B 5/7267A61B 6/481G16H 50/20G16H 30/20G16H 30/40

PatentIndex Score

Cited by

References

Claims

Abstract

According to various examples, pharmacokinetics of a contrast medium through a cardiovascular system of a patient are estimated. This is achieved using analytical model(s) such as a compartment model or a linear time invariant model, and/or one or more neural network algorithms. An angiographic imaging protocol may be configured based on the estimate of the pharmacokinetics.

Claims

exact text as granted — not AI-modified

What is claimed is: 
     
         1 . A computer-implemented method, comprising:
 training a neural network algorithm to estimate, based on a measurement indicative of a reference contrast medium bolus, pharmacokinetics of a contrast medium through a cardiovascular system of a patient, the training being dependent on at least one loss; and   configuring an angiographic imaging protocol based on an estimate of the pharmacokinetics of the contrast medium obtained from the neural network algorithm, wherein
 the at least one loss includes a loss determined based on a compartment model of the cardiovascular system, the compartment model being based on a fragmentation of the cardiovascular system into multiple compartments and associated differential equations defining a change in a concentration of the contrast medium as a function of time and a concentration of the contrast medium at a respective compartment. 
   
     
     
         2 . The computer-implemented method of  claim 1 ,
 wherein the loss that is determined based on the compartment model is based on a comparison between time and spatial derivatives of a prediction of the neural network algorithm for the pharmacokinetics of the contrast medium and further time and spatial derivatives of the concentration of the contrast medium as defined by the associated differential equations of the compartment model.   
     
     
         3 . A computer-implemented method, comprising:
 training a neural network algorithm to estimate, based on a measurement indicative of a reference contrast medium bolus, pharmacokinetics of a contrast medium through a cardiovascular system of a patient, the training being dependent on at least one loss; and   configuring an angiographic imaging protocol based on an estimate of the pharmacokinetics of the contrast medium obtained from the neural network algorithm, wherein
 the at least one loss includes a loss determined based on a linear time invariant model of the cardiovascular system, the linear time invariant model being based on a response function of a vessel of the cardiovascular system with respect to a time-dependent inflow of the contrast medium. 
   
     
     
         4 . The computer-implemented method of  claim 1 ,
 wherein the measurement includes time series data that includes samples that are irregularly spaced in the time domain, and   wherein the neural network algorithm includes a Neural Ordinary Differential Equations algorithm.   
     
     
         5 . The computer-implemented method of  claim 1 ,
 wherein the at least one loss includes multiple losses, and   wherein the method further includes weighting the multiple losses.   
     
     
         6 . A computer-implemented method, comprising:
 obtaining an estimate of pharmacokinetics of a contrast medium through a cardiovascular system of a patient, the estimate being obtained from a neural network algorithm that obtains, as input, at least a three-dimensional image of the patient; and   configuring an angiographic imaging protocol based on the estimate of the pharmacokinetics of the contrast medium.   
     
     
         7 . A computer-implemented method, comprising:
 determining one or more parameter values of at least one free parameter of an analytical model that estimates pharmacokinetics of a contrast medium through a cardiovascular system of a patient using an optimization algorithm, the optimization algorithm operating based on an input including a measurement indicative of a reference contrast medium bolus; and   configuring an angiographic imaging protocol based on an estimate of the pharmacokinetics of the contrast medium obtained from the analytical model.   
     
     
         8 . The computer-implemented method of  claim 7 , wherein the at least one free parameter includes a time-dependent in-flow concentration of a contrast-medium measurement bolus. 
     
     
         9 . A computer-implemented method, comprising:
 obtaining an estimate of pharmacokinetics of a contrast medium through a cardiovascular system of a patient, the estimate being obtained from a neural network algorithm that includes an operator layer implementing an analytical model that also estimates the pharmacokinetics of the contrast medium through the cardiovascular system of the patient; and   configuring an angiographic imaging protocol based on the estimate of the pharmacokinetics of the contrast medium obtained from the neural network algorithm.   
     
     
         10 . The computer-implemented method of  claim 9 ,
 wherein the analytical model includes at least one free parameter; and   wherein the method further includes determining one or more parameter values of the at least one free parameter using an optimization algorithm, the optimization algorithm operating based on an input including a measurement indicative of a reference contrast medium bolus.   
     
     
         11 . A computer-implemented method, comprising:
 determining multiple pre-estimates of pharmacokinetics of a contrast medium through a cardiovascular system of a patient based on at least one of multiple models or neural network algorithms;   determining an estimate of the pharmacokinetics based on a weighted combination of the multiple pre-estimates; and   configuring an angiographic imaging protocol based on the estimate of the pharmacokinetics.   
     
     
         12 . The computer-implemented method of  claim 11 , further comprising:
 determining weights of the weighted combination based on ensemble learning.   
     
     
         13 . A method executed by at least one processor of a computing device upon loading program code from a memory of the computing device, the method comprising:
 determining one or more estimates of pharmacokinetics of a contrast medium through a cardiovascular system of a patient based on one or more analytical models and one or more neural network algorithms; and   configuring an angiographic imaging protocol based on the one or more estimates of pharmacokinetics of the contrast medium.   
     
     
         14 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 1 .   
     
     
         15 . A non-transitory computer-readable medium storing computer-executable instructions that, when executed by at least one processor at a computing device, cause the computing device to perform the method of  claim 1 . 
     
     
         16 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 3 .   
     
     
         17 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 6 .   
     
     
         18 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 7 .   
     
     
         19 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 9 .   
     
     
         20 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 11 .   
     
     
         21 . A computing device comprising:
 at least one processor and a memory, the at least one processor being configured to load program code from the memory and to execute the program code, wherein the at least one processor, upon executing the program code, is configured to cause the computing device to perform the method of  claim 13 .   
     
     
         22 . The computer-implemented method of  claim 2 ,
 wherein the measurement includes time series data that includes samples that are irregularly spaced in the time domain, and   wherein the neural network algorithm includes a Neural Ordinary Differential Equations algorithm.   
     
     
         23 . The computer-implemented method of  claim 2 ,
 wherein the at least one loss includes multiple losses, and   wherein the method further includes weighting the multiple losses.   
     
     
         24 . The computer-implemented method of  claim 3 ,
 wherein the measurement includes time series data that includes samples that are irregularly spaced in the time domain, and   wherein the neural network algorithm includes a Neural Ordinary Differential Equations algorithm.   
     
     
         25 . The computer-implemented method of  claim 3 ,
 wherein the at least one loss includes multiple losses, and   wherein the method further includes weighting the multiple losses.

Join the waitlist — get patent alerts

Track US2024138793A1 — get alerts on status changes and closely related new filings.

We store only your email — no account needed. See our privacy policy.